diff --git a/kmath-noa/README.md b/kmath-noa/README.md
index 1238dc885..9ca1933f6 100644
--- a/kmath-noa/README.md
+++ b/kmath-noa/README.md
@@ -10,5 +10,22 @@ which covers a wide set of applications from particle physics
 simulations to deep learning and general differentiable programs
 written on top of `AutoGrad` & `ATen`.
 
-Currently, the native artifacts support only `GNU` and 
-`CUDA` for GPU acceleration. 
\ No newline at end of file
+## Installation
+
+Currently, to build native artifacts, we support only 
+the [GNU](https://gcc.gnu.org/) toolchain. For `GPU` kernels, we require a compatible 
+[CUDA](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html)
+installation. If you are on Windows, we recommend setting up
+everything on [WSL](https://docs.nvidia.com/cuda/wsl-user-guide/index.html).
+
+To install the library, you have to publish 
+locally `kmath-core`, `kmath-tensors` with `kmath-noa`:
+
+```
+./gradlew -q :kmath-core:publishToMavenLocal :kmath-tensors:publishToMavenLocal :kmath-noa:publishToMavenLocal
+```
+
+This builds `jtorch` a JNI wrapper for `NOA/LibTorch`, placed inside:
+
+`~/.konan/third-party/kmath-noa-<version>/cpp-build`
+
diff --git a/kmath-noa/build.gradle.kts b/kmath-noa/build.gradle.kts
index b2770df1e..34c7b730b 100644
--- a/kmath-noa/build.gradle.kts
+++ b/kmath-noa/build.gradle.kts
@@ -18,8 +18,8 @@ dependencies {
     api(project(":kmath-tensors"))
 }
 
-val home = System.getProperty("user.home")
-val javaHome = System.getProperty("java.home")
+val home: String = System.getProperty("user.home")
+val javaHome: String = System.getProperty("java.home")
 val thirdPartyDir = "$home/.konan/third-party/kmath-noa-${project.property("version")}"
 val cppBuildDir = "$thirdPartyDir/cpp-build"
 val cppSources = projectDir.resolve("src/main/cpp")